The Number Line Effect Reflects Top-Down Control
Ristic, Jelena, Wright, Alissa, Kingstone, Alan, Psychonomic Bulletin & Review
Recent evidence indicates that central directional stimuli, such as eyes and arrows, trigger rapid, reflexive shifts of spatial attention. A study by Fischer, Castel, Dodd, and Pratt (2003) suggested that a similar effect might also apply to central numbers, as if a digit's meaning causes attention to be oriented to its relative position on a left-to-right mental number line. However, unlike central eyes and arrows, the orienting effect for central digits emerges slowly, suggesting that top-down endogenous processes may be mediating this effect. Here, we report a series of three experiments that strongly support this hypothesis. Experiment 1 replicated Fischer et al.'s left-to-right number line effect. Experiment 2 showed that this effect could be completely reversed by merely asking participants to imagine a number line running from right to left. Experiment 3 showed that a left-to-right number line effect could be abolished by presenting targets above and below central fixation, as well as to the left and right of center. Experiment 3 also showed that other mental sets, such as imagining a clock, result in attention's being oriented in accordance with where the central digits are represented on a clock face. Together, these data indicate that the spatial representations and attentional orienting related to the perception of digits are both fragile and flexible and depend critically on the top-down spatial mental sets adopted by individuals.
Human attention is normally conceived as a limited capacity process that can be controlled in either an exogenous, reflexive manner or an endogenous, volitional manner (Broadbent, 1971; Posner, 1978). For several decades, the attention-cuing paradigm has provided a simple methodology for engaging and measuring these two forms of orienting (Posner, 1980; Posner, Snyder, & Davidson, 1980).
Typically, reflexive attention has been triggered by a peripheral flash that does not predict where a visual target will appear. This attentional cue is followed by a target, demanding a manual detection response, appearing either at the cued location or at a noncued location. The usual result is that if the cue-target stimulus onset asynchrony (SOA) is less than 300 msec, reaction time (RT) is shorter when a target appears at a cued location versus a noncued location. Because attended items are processed more efficiently than nonattended items, the facilitation effect at the cued location is taken as evidence that the abrupt onset attracted attention to its location. And because the facilitation effect emerged rapidly in response to a cue that did not predict where the target would appear, the attention effect is considered to be reflexive in nature. For SOAs greater than 300 msec, RT becomes longer for targets appearing at the cued location than for those appearing at the noncued locations, reflecting the inhibition-of-return phenomenon. This inhibition effect at the cued location is taken as evidence that attention was withdrawn from the cued location and is inhibited in returning there (Posner &Cohen, 1984).
The cuing methodology for volitional orienting is similar to the above, with two important exceptions. First, instead of an attentional cue being flashed in the periphery, a central directional cue, usually an arrow, points toward the cued location. second, the arrow cue predicts where the target is most likely to appear. The usual result when this central predictive cuing methodology is used is that an RT advantage emerges relatively slowly at the cued location and then persists across the long cue-target SOAs, since there is little reason to shift attention away from where the target is most likely to appear (e.g., Jonides, 1981).
Recently, a number of studies have shown that central directional cues do not need to be spatially predictive in order to induce a shift in spatial attention to the cued location. Friesen and Kingstone (1998) reported that a simple, spatially nonpredictive central schematic face looking left or right triggers a shift in attention to the gazed-at (cued) location. …